"Modern Family? The Gendered Effects of Marriage and Chilbbearing on Voter Turnout" REPLICATION PACKAGE
By Giorgio Bellettini, Carlotta Berti Ceroni, Enrico Cantoni, Chiara Monfardini, and Jerome Schafer

FILE DESCRIPTION:		REPLICATION PACKAGE README FILE
AUTHOR:						ENRICO CANTONI
LAST EDITED ON:		NOVEMBER 8, 2022


**********************************************************************************************************************************


*** SOFTWARE REQUIREMENTS ***

Stata version 16.1 (64-bit) or higher is required to run this replication package.


**********************************************************************************************************************************


*** CODE REQUIREMENTS ***

The code of this replication package is "self-standing". That is, it contains the ado files for all non-Stata embedded commands used throughout the code. This ensures the outcome of the replication does not depend on the version of the Stata commands downloaded from SSC that are installed on the local machine.


Yet, to run the cleaning and analysis code, the user must change the global macro "root" on line 20 of the master cleaning and analysis files, respectively, (i.e., clean.do and analysis.do) to reflect the location of the replication folder on the user�s machine.


**********************************************************************************************************************************


*** FOLDER STRUCTURE ***

The main replication folder contains three subfolders:

1. Code
Contains all analysis do-files (in the Analysis subfolder), data cleaning do-files (in the Cleaning subfolder), the ado-files of the necessary user-written Stata command (in the ado subfolder), and the global.do file specifying all global macros used in other files.
Since we included all necessary adofiles, the user will not need to install any external Stata package to replicate our results.
The inclusion of the adofiles thus ensures that the commands used for replication are identical to and will work exactly as the commands we used for our analyses.
In turn, this guarantees that results will not change because of differences in the version of the external Stata packages used by the replicating user vs. in the original analyses.

globals.do
Specifies all global macros used in other scripts

Code/ado/personal
Contains the adofiles underlying the Stata commands we wrote ourselves.

Code/ado/plus
Contains the adofiles, help files, and ancillary files underlying the Stata commands we installed from external archives (e.g., SSC).

Code/Analysis/analysis.do
Is the master analysis file.

!!!!! NOTE: All analysis files must be run through this file. This file includes local "switches" allowing users to run only certain parts of the analysis code. !!!!!

!!!!! NOTE: to run the analysis code, change the global macro "root" on line 20 of the master analysis file to reflect the location of the replication folder on the user�s machine. !!!!!

Code/Analysis/analysis_descriptive.do
Computes untabulated descriptive statistics mentioned in the text, but not included in any table of figure.

Code/Analysis/analysis_graphs_descriptive.do
Creates Figures 1, 2, A1, A2, and A4.

Code/Analysis/analysis_graphs_event.do
Creates Figures 3, 5, A5, A6, A7, A8, and A9.

Code/Analysis/analysis_graphs_kids.do
Creates Figure 4.

Code/Analysis/analysis_tables_bologna.do
Creates Tables 2, 3, A3, A4, A5, A6, and A7.

Code/Analysis/analysis_tables_summary.do
Creates Tables 1, A1, and A2.

Code/Cleaning/clean.do
Is the master cleaning file.

!!!!! NOTE: All cleaning files must be run through this file. This file includes local "switches" allowing users to run only certain parts of the cleaning code. !!!!!

!!!!! NOTE: to run the cleaning code, change the global macro "root" on line 20 of the master cleaning file to reflect the location of the replication folder on the user�s machine. !!!!!

Code/Cleaning/clean_CPS.do
Cleans the Current Population Survey (CPS) data.


2. Data
Contains all the clean and raw data.

Data/Clean
Contains the clean data ready for analyses.

Data/Clean/analysis_clean.dta
Is the anonymous voter-level income dataset from the Municipality of Bologna.

Data/Clean/CPS_clean.dta
Is the clean CPS dataset.

Data/Raw
Contains the original raw files.

Data/Raw/Cattaneo/Database1994_2006.dta
Is the original file of the Istituto Cattaneo voter turnout data (used for Figure A2).

Data/Raw/CPS/cps_00003.dta
Is the original file for the CPS data (used for Figure 1a).

Data/Raw/Interno/storico_amministratori_comuni31122012.csv
Contains information (e.g., gender) on all public officials (including city councillors) working for Italian municipalities as of December 31, 2012.

Data/Raw/ISTAT/lfp_byprovinceyear_2004to2019.csv
Contains figures on labor force participation by gender and education and year at the province level, 2004-2019.

Data/Raw/ISTAT/nurseryschools_bycityyear.csv
Contains figures on nursery school seats by province, year, and school type, 2013-2017.

Data/Raw/ISTAT/pop_bycity_2013.csv
Contains figures on the resident population by municipality, gender, and marital status in 2013.

Data/Raw/KostelkaBlais/Base_dataset.dta
Is Kostelka and Blais (World Politics, 2021)'s main replication dataset (used for Figure A1).

Data/Raw/Unibo/values.xlsx
Is an ancillary file used to parsimoniously create new variables (using the Stata command "fix_excel").

Data/Raw/Unibo/variables.xlsx
Is an ancillary file used to parsimoniously bulk rename and relabel variables in Stata (using the Stata command "renamefrom").


3. Output
Contains all the figures and tables created during the analyses.

Output/Graphs
Contains all the figures in both gph and pdf format.

Output/Graphs/Figure_1_a.gph
Output/Graphs/Figure_1_a.pdf
Voter turnout by age and gender in the U.S. based on CPS data (Figure 1a).

Output/Graphs/Figure_1_b.gph
Output/Graphs/Figure_1_b.pdf
Voter turnout by age and gender in Bologna based on our own proprietary data (Figure 1b).

Output/Graphs/Figure_2_a.gph
Output/Graphs/Figure_2_a.pdf
Gender-specific average turnout by income among never-married voters (Figure 2a).

Output/Graphs/Figure_2_b.gph
Output/Graphs/Figure_2_b.pdf
Gender-specific average turnout by income among married voters (Figure 2b).

Output/Graphs/Figure_3_a.gph
Output/Graphs/Figure_3_a.pdf
Gender-specific event-study estimates of the impact of marriage on turnout (Figure 3a).

Output/Graphs/Figure_3_b.gph
Output/Graphs/Figure_3_b.pdf
Event-study estimates of the impact of marriage on the female-minus-male turnout gap (Figure 3b).

Output/Graphs/Figure_4_a.gph
Output/Graphs/Figure_4_a.pdf
Gender-specific estimates of the effect of children on parental turnout by children's age (Figure 4a).

Output/Graphs/Figure_4_b.gph
Output/Graphs/Figure_4_b.pdf
Estimates of the impact of children of different ages on the female-minus-male parental turnout gap (Figure 4b).

Output/Graphs/Figure_5_a.gph
Output/Graphs/Figure_5_a.pdf
Event-study estimates of the impact of children on parental turnout (Figure 5a).

Output/Graphs/Figure_5_b.gph
Output/Graphs/Figure_5_b.pdf
Event-study estimates of the impact of children on the female-minus-male turnout gap (Figure 5b).

Output/Graphs/Figure_A1.gph
Output/Graphs/Figure_A1.pdf
Evolution of voter turnout in Italy and other democratic countries (based on data from Kostelka and Blais', 2021; Figure A1).

Output/Graphs/Figure_A2.gph
Output/Graphs/Figure_A2.pdf
Voter turnout in Italy and in Emilia-Romagna by voter gender, 1994-2006, based on Istituto Cattaneo administrative voter-level panel data (Figure A2).

Output/Graphs/Figure_A3.gph
Output/Graphs/Figure_A3.pdf
Sample data rows (Figure A3).

Output/Graphs/Figure_A4.gph
Output/Graphs/Figure_A4.pdf
Lowess regression of voter turnout on OECD-modified household income by voter gender and age range (Figure A4).

Output/Graphs/Figure_A5.gph
Output/Graphs/Figure_A5.pdf
Event study of the impact of marriage on turnout, after rescaling estimates by gender-specific mean turnout in the last pre-marriage election (Figure A5).

Output/Graphs/Figure_A6.gph
Output/Graphs/Figure_A6.pdf
Event study of the impact of children on turnout, after rescaling estimates by gender-specific mean turnout in the last pre-children election (Figure A6).

Output/Graphs/Figure_A7_a.gph
Output/Graphs/Figure_A7_a.pdf
Event-study estimates of the impact of children on parental turnout, restricting the sample to voters who, in the first post-children election, make less than 70% of their income in the last pre-children election (Figure A7a).

Output/Graphs/Figure_A7_b.gph
Output/Graphs/Figure_A7_b.pdf
Event-study estimates of the impact of children on the female-minus-male turnout gap, restricting the sample to voters who, in the first post-children election, make less than 70% of their income in the last pre-children election (Figure A7b).

Output/Graphs/Figure_A8_a.gph
Output/Graphs/Figure_A8_a.pdf
Event-study estimates of the impact of children on parental turnout, restricting the sample to voters who, in the first post-children election, make 70% or more of their income in the last pre-children election (Figure A8a).

Output/Graphs/Figure_A8_b.gph
Output/Graphs/Figure_A8_b.pdf
Event-study estimates of the impact of children on the female-minus-male turnout gap, restricting the sample to voters who, in the first post-children election, make 70% or more of their income in the last pre-children election (Figure A8b).

Output/Graphs/Figure_A9_a.gph
Output/Graphs/Figure_A9_a.pdf
Gender-specific event-study estimates of the impact of cohabitation on turnout (Figure A9a).

Output/Graphs/Figure_A9_b.gph
Output/Graphs/Figure_A9_b.pdf
Event-study estimates of the impact of cohabitation on the female-minus-male turnout gap (Figure A9b).


Output/Tables
Contains all the tables.
Stata exports each table as a separate tab in the spreadsheet tables_raw.xlsx
This file only contains the table content; i.e., it displays no row or column headers, among others.
Instead of looking at tables_raw.xlsx, users should look at tables_shell_replication.xlsm, which contains easy-to-read tables shells in a format that is virtually identical to the tables included in the paper (or in the appendix).
The content of these tables is automatically loaded from tables_raw.xlsx; hence, if tables_raw.xlsx is edited, tables_shell_replication.xlsm is automatically updated.